Mutually Reinforced Multicomponent Polysaccharide Networks...Mutually Reinforced Multicomponent Polysaccharide Networks Laura L. Hyland,1 Marc B. Taraban,1 Boualem Hammouda,2 Y. Bruce

Mutually Reinforced Multicomponent Polysaccharide NetworksLaura L. Hyland,1 Marc B. Taraban,1 Boualem Hammouda,2 Y. Bruce Yu1,31 Fischell Department of Bioengineering, University of Maryland, College Park, MD 20742

2 NIST Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, MD 20899

3 Department of Pharmaceutical Sciences, School of Pharmacy, University of Maryland, Baltimore, MD 21201

Received 2 March 2011; revised 26 May 2011; accepted 31 May 2011

Published online 22 June 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.21687

This article was originally published online as an accepted

preprint. The ‘‘Published Online’’ date corresponds to the

preprint version. You can request a copy of the preprint by

emailing the Biopolymers editorial office at biopolymers@wiley.

com

INTRODUCTION

Chitosan has become one of the most commonly uti-

lized biopolymers in biomaterials research. This cati-

onic polysaccharide has many attractive qualities and

is abundantly found in nature.1 Chitosan has been

widely studied for tissue engineering applications

because of its biocompatibility and biodegradability. Its deg-

radation products are glucosamine and N-acetyl glucosa-

mine, amino sugars naturally found in the human body. The

hydrophilic surface of chitosan has been shown to promote

cell adhesion, proliferation, and differentiation.2-4 Chitosan

Mutually Reinforced Multicomponent Polysaccharide Networks

Additional Supporting Information may be found in the online version of this article.Correspondence to: Y. Bruce Yu; e-mail: [email protected]

ABSTRACT:

Networks made from chitosan and alginate have been

utilized as prospective tissue engineering scaffolds due to

material biocompatibility and degradability. Calcium

(Ca21) is often added to these networks as a modifier for

mechanical strength enhancement. In this work, we

examined changes in the bulk material properties of

different concentrations of chitosan/alginate mixtures (2,

3, or 5% w/w) upon adding another modifier,

chondroitin. We further examined how material

properties depend on the order the modifiers, Ca21 and

chondroitin, were added. It was found that the addition

of chondroitin significantly increased the mechanical

strength of chitosan/alginate networks. Highest elastic

moduli were obtained from samples made with mass

fractions of 5% chitosan and alginate, modified by

chondroitin first and then Ca21. The elastic moduli in

dry and hydrated states were (4.416 0.52) MPa and

(0.11 6 0.01) MPa, respectively. Network porosity and

density were slightly dependent on total polysaccharide

concentration. Average pore size was slightly larger in

samples modified by Ca21 first and then chondroitin and

in samples made with 3% starting mass fractions. Here,

small-angle neutron scattering (SANS) was utilized to

examine mesh size of the fibrous networks, mass-fractal

parameters and average dimensions of the fiber cross-

sections prior to freeze-drying. These studies revealed that

addition of Ca21 and chondroitin modifiers increased

fiber compactness and thickness, respectively. Together

these findings are consistent with improved network

mechanical properties of the freeze-dried materials.

# 2011 Wiley Periodicals, Inc. Biopolymers 95: 840–851,

2011.

Keywords: chitosan; alginate; chondroitin; compression-

tensile tester; freeze-dry; correlation length; fractal

dimensions; scanning electron microscopy (SEM); small-

angle neutron scattering (SANS)

Contract grant sponsor: NIH

Contract grant number: EB004416

Contract grant sponsor: National Science Foundation

Contract grant number: DMR-0944772

Contract grant sponsor: Maryland Technology Development Corporation

(TEDCO)

VVC 2011 Wiley Periodicals, Inc.

840 Biopolymers Volume 95 / Number 12

is also versatile; it is easily moldable and has many functional

groups that can be modified to tune material properties.5

However, by itself chitosan is mechanically weak and swells

to disassembly in aqueous environments.6

Alginate is an anionic polysaccharide that can electrostati-

cally interact with cationic chitosan.7 Upon interaction, algi-

nate and chitosan form fibers which create a gel-like, solid

material. This material can be freeze-dried and mechanically

tested. Like chitosan, alginate is a widely used biocompatible

polymer, which is known to support the proliferation of cells

both in vitro and in vivo.8,9 However, on its own alginate is a

viscous, weak material. When used as a component in scaf-

folds, alginate is often modified with divalent cations like

Ca21 to create a strong gel with a characteristic egg box

structure.10

A number of networks have been made using combina-

tions of chitosan and alginate with Ca21 as a modifier. These

materials were made by combining and freeze-drying the

mixtures to create novel biomaterials. Uses for these net-

works include bone replacements,11 liver replacements,12 and

medicated wound dressings.13 These studies have examined

chitosan-alginate networks at low polysaccharide mass frac-

tions (0.05%–2.4%), but give valuable insight about network

characteristics such as tunability and cell compatibility. The

strongest chitosan-alginate networks to date were made with

a mass fraction of 2.4% chitosan and a mass fraction of 2.4%

alginate and had a dry compressive elastic modulus of (2.56

6 0.41) MPa.11 These networks could support osteoblast

attachment, proliferation and also calcium deposition. Here,

the potential of the chitosan-alginate networks as load-bear-

ing biomaterials was demonstrated. However, these data

lacked the important mechanical characteristics in the bio-

logically relevant hydrated state. Therefore, more studies are

necessary.

In addition to alginate, chitosan can interact with glycos-

aminoglycans (GAGs) which are also anionic polysaccha-

rides. GAGs are valuable because they can facilitate the

migration and proliferation of progenitor cells promoting

tissue regeneration.17,18 Chondroitin sulfate is one kind of

commercially available GAG. We found that this anionic

polysaccharide creates fibrous, elastic networks with the cati-

onic chitosan upon mixing. Chitosan-chondroitin networks

have been used for the controlled release of platelet-derived

growth factor for bone regeneration. In vitro drug release

could be controlled by adjusting the ratio of chitosan to

chondroitin.19

To improve the mechanical properties of the scaffolds, we

hypothesize that the incorporation of chondroitin as a sec-

ond modifier into the chitosan-alginate-Ca21 network could

increase electrostatic interactions and improve its overall

strength and flexibility. Further, we examined the effect of

the order of adding each of the two modifiers, Ca21 and

chondroitin, on the mechanical strength of the network. To

this end, three types of networks were prepared: type A,

which are chitosan/alginate networks with Ca21 as the sole

modifier; type B, which are chitosan/alginate networks with

Ca21 added as the 1st modifier and chondroitin added as the

2nd modifier; type C, which are chitosan/alginate networks

with chondroitin added as the 1st modifier and Ca21 added

as the 2nd modifier. The resulting freeze-dried networks were

tested for their compression and tensile strengths.

To promote cell proliferation and migration in vivo, net-

works should have high porosity, suitable and non-uniform

pore size, and highly interconnected pore structure in addi-

tion to biocompatibility and biodegradability.14-16 Therefore,

network porosity, density and pore size of the freeze-dried

materials were examined to determine the effect of the poly-

saccharide content, the addition of a 2nd modifier, as well as

the addition order of the two modifiers, on these properties.

It would be reasonable to suggest that mechanical strength

of the freeze-dried polysaccharide scaffolds would depend on

the structural characteristics of the polysaccharide networks

formed in solution when mixing the components prior to

freeze-drying. Therefore, to aid our understanding of the

interactions between the modifiers and the chitosan/alginate

scaffold, small-angle neutron scattering (SANS) in solution

was used to investigate the impact of the addition of modi-

fiers Ca21 and chondroitin individually on the structural fea-

tures of the chitosan/alginate network. This approach has

allowed us to trace how the structural features at the level of

individual fiber and the polysaccharide network as a whole

are translated into the bulk material properties upon freeze-

drying.

MATERIALS AND METHODS

Preparation of Networks for Mechanical

and Imaging StudiesLow molecular weight chitosan (50–190 kDa, Sigma-Aldrich), al-

ginic acid sodium salt (350–450 kDa, Acros Organics), bovine chon-

droitin sulfate sodium salt (� 20 kDa, Pfaltz & Bauer), hydrochloric

acid (HCl, VWR), ammonium hydroxide (NH4OH, Mallinckrodt

Baker), ethanol (EMD) and calcium chloride dihydrate (CaCl2 �2H2O, Mallinckrodt Baker) were used as purchased.

Solutions of mass fractions 2, 3, and 5% chitosan were prepared

in a mass fraction of 2% HCl in ultrapure water (18.2 MOhm, 2 lmcellulose filter) while solutions of mass fractions 2%, 3% and 5% al-

ginate were prepared in a mass fraction of 2% NH4OH. Mass frac-

tions of 1% CaCl2 and 2% chondroitin solutions were prepared in

ultrapure water (18.2 MOhm, 2 lm cellulose filter). To prepareContract grant sponsor:

Mutually Reinforced Multicomponent Polysaccharide Networks 841

Biopolymers

sample type A (Figure 1), alginate and chitosan were mixed together

at equal concentrations and equal volumes. Type A samples were

made at three polysaccharide concentrations by mixing mass frac-

tions of 2% chitosan with mass fractions of 2% alginate, mass frac-

tions of 3% chitosan with mass fractions of 3% alginate and mass

fractions of 5% chitosan with mass fractions of 5% alginate, with

the resulting samples labeled as 2A, 3A, and 5A, respectively. The

electrostatic interactions between chitosan and alginate upon mix-

ing resulted in fibrous, gel-like materials. After chitosan and alginate

mixing, the 1st modifier, 1% mass fraction CaCl2 solution, was

added at a volume ratio of 10:1 chitosan-alginate:CaCl2 for all type

A samples. The samples were then placed in a 2208C freezer over-

night and then lyophilized. After lyophilization, dried type A sam-

ples were soaked in ultrapure water at room temperature for 30

min. Samples were frozen at 2208C and lyophilized again. At this

point, they were ready for testing. Type B samples (Figure 1) were

also made using the same chitosan and alginate mixing concentra-

tions. Again, the first modifier CaCl2 was added. Dried type B sam-

ples were then soaked in the 2nd modifier 2% mass fraction chon-

droitin at room temperature for 30 min. The soaked samples were

frozen at2208C and lyophilized once again. Type C samples (Figure

1) were made by adding 2% mass fraction chondroitin as the first

modifier at a volume ratio of 6:1 chitosan-alginate:chondroitin and

1% mass fraction CaCl2 as the 2nd modifier. For compressive test-

ing, the dried samples were sliced into 12 mm thick dry cylinders.

FIGURE 1 Procedures for making the three types of networks, A, B and C. Each network under-

went lyophilization twice. x% chitosan was mixed with x% alginate in a 1:1 volume ratio. Type A &

B samples were made by adding Ca21 to the chitosan-alginate mixture at a volume ratio of 10:1

chitosan-alginate:CaCl2. Type A & B samples were lyophilized and then soaked in ultrapure H2O

and a mass fraction of 2% chondroitin respectively. Type C samples were made by adding chon-

droitin to the chitosan-alginate mixture at a volume ratio of 6:1 chitosan-alginate:chondroitin.

Type C samples were lyophilized and soaked in a mass fraction of 1% CaCl2.

FIGURE 2 A representative image of the cylinder-shaped version

of the freeze-dried samples.

842 Hyland et al.

Biopolymers

The diameter for each dry cylinder was � 20 mm. For tensile test-

ing, the same sample-making procedure was used except samples

were sliced into rectangular plates, 10 mm wide and 40 mm long

and 2–3 mm thick. Finished samples were completely dry, solid

materials. Figure 2 shows a representative image of the cylinder-

shaped version of these freeze-dried samples.

Mechanical TestingMechanical strength of the freeze-dried networks was assessed

using a Tensilon tensile-compressive tester (RTF-1310, Orientec,

Japan) with a 50 N load cell. For compressive testing, the guide-

lines for mechanical testing from ASTM D5024-95a were used as

described.11,20 Briefly, the freeze-dried samples were hydrated to

saturation and compressed to 30% of their original thicknesses

with a constant crosshead speed of 0.4 mm/min. For tensile test-

ing, rectangular freeze-dried networks were hydrated to saturation

and elongated until rupture at a crosshead speed of 6.0 mm/min.21

Elastic moduli from compressive tests were calculated using the

slopes of their respective stress–strain curves. To obtain the most

realistic mechanical values, samples were tested in a hydrated state.

However, the strongest sample (5C) was compressed in a dry state

in order to compare with other reported chitosan-alginate strength

values.11 Ultimate tensile strength was calculated by dividing the

maximum load value by the material cross-section. The strongest

sample (5C) was also put under tension in a dry state to determine

the difference between dry and hydrated states. Five samples were

used for each mechanical test. Mechanical testing results are pre-

sented as the average of five sample tests with the standard devia-

tion reported as the error.

Scanning Electron MicroscopyTotally, 10 mm 3 10 mm pieces of each dried sample were exam-

ined using Scanning Electron Microscopy (SEM, Hitachi SU-70).

Samples were placed on an SEM sample holder and coated with a

thin layer of gold (�30 nm) using a Sputter Coater (Anatech

Hummer X). Average pore diameters of the networks were deter-

mined using the NIH image analysis program, ImageJ.22 Six images

from each sample were taken for analysis of the entire sample sur-

face. Every pore was measured in all images.

Material Porosity and DensityA liquid displacement method described by Zhang et al. was modi-

fied and used to determine the polysaccharide network porosity and

density.20 Dried samples of dimensions 7 mm 3 7 mm 3 7 mm

were weighed (W) and then placed in a known volume of liquid

(V1). Air was evacuated from the samples followed by repressuriza-

tion to insure maximum liquid saturation. The residual pressure

here was close to 20 Torr. Air evacuation was done using a modified

graduated cylinder, fitted with an attachment for vacuum pumping.

The volume of the liquid including the saturated network (V2) was

measured. The saturated network was then removed and the

remaining liquid volume (V3) was measured. The original method

used ethanol to determine porosity because it does not cause net-

work swelling. However, we found ethanol evaporation to be a

problem during air evacuation. Instead, heptane was used as the dis-

placement liquid. Heptane did not have noticeable evaporation dur-

ing air evacuation and did not affect network swelling. The density

(q) and porosity (e) of the networks were then calculated using the

following equations.

q ¼ weight of dry network

volume of solvated network¼ W

V2 � V3

ð1Þ

e ¼ volume of liquid in solvated network

volume of solvated network¼ V1 � V3

V2 � V3

ð2Þ

Preparation of Networks for Small-Angle Neutron

Scattering (SANS) StudyChitosan, alginate, chondroitin and calcium chloride solutions were

prepared in D2O to enable adequate contrast between the hydrogen-

rich networks and the solvent. Solutions of mass fraction 2% chito-

san were made in D2O containing a mass fraction of 2% HCl and

solutions of mass fraction 2% alginate were made in D2O contain-

ing a mass fraction of 2% NH4OH. Solutions of a mass fraction of

1% chondroitin and 0.5% CaCl2 were each made in D2O. Five sam-

ples were prepared for measurement (Table I). The calcium contain-

ing sample was made by mixing a mass fraction of 2% chitosan with

a mass fraction of 2% alginate in equal volumes and then calcium

was added at a volume ratio of 10:1 chitosan-alginate:CaCl2. The

chondroitin containing sample was made using the same chitosan

Table I Structural Data From SANS Analysis

Samples lc (A) D B Rc (A)

2% chitosan 147 6 8 3.0 6 0.3 — 83

2% alginate 245 6 9 2.6 6 0.3 — 88

2% chitosan1 2% alginate 134 6 5 2.9 6 0.2 4.6 3 1024 109

2% chitosan 1 2% alginate 1 0.25% Ca21 120 6 5 2.8 6 0.2 1.0 3 1024 92

2% chitosan 1 2% alginate 1 0.5% chondroitin 149 6 8 3.0 6 0.2 3.9 3 1024 126

Correlation length (lc), mass-fractal (d), mass-fractal prefactor (B), and radius of gyration of the cross-section (Rc), were analyzed for chitosan/alginate

samples. Each mixture was made with equal volumes of a mass fraction of 2% chitosan and a mass fraction of 2% alginate. The calcium-containing sample

was made by adding Ca21 to the chitosan/alginate mixture at a volume ratio of 10:1 chitosan-alginate:CaCl2 (0.25% CaCl2). The chondroitin-containing sam-

ple was made by adding chondroitin to the chitosan/alginate mixture at a volume ratio of 6:1 chitosan/alginate/chondroitin (0.5% chondroitin). The B for

chitosan and alginate could not be calculated due to low scattering values.


Biopolymers

and alginate mixture and chondroitin was added at a volume ratio

of 6:1 chitosan-alginate:chondroitin. Mixtures were prepared within

titanium sample cells with 30 mm diameter quartz windows and a

1-mm path length. Samples were prepared within 12 h of measure-

ments. Of note, the samples for SANS experiments were not freeze-

dried as opposed to the samples used for SEM, and mechanical, po-

rosity and density studies. We have performed SANS experiments

with polysaccharide networks in solution before they were freeze-

dried in an attempt to get an insight on how the structural charac-

teristics of the polysaccharide networks at the nanoscale level (or

the level of individual fiber) are further translated into the bulk ma-

terial properties. Because of dimensional hindrances of 1-mm

quartz-titanium sample cell used in SANS studies, freeze-dried sam-

ples could not be loaded. We were also limited to lower concentra-

tions of polysaccharides which contained only one modifier for

each. High viscosity of concentrated solutions as well as the diffu-

sion limitations for modifiers in the restricted environment of the

sample cell hampered the extension of our experiments to wider

concentration ranges and the addition of a second modifier. How-

ever, despite the above limitations, SANS studies can provide solid

support for the results of bulk material testing and in some sense

could serve as a basis for explanation of the observed material prop-

erties.

SANS Structural AnalysisStructures of the networks listed in Table I were investigated using

the 30 m SANS instrument (NG-3)23 at the National Institute of

Standards and Technology (NIST). Neutrons at k 5 6 A with a

wavelength spread (Dk/k) of 0.14 were detected on a 64 cm 3 64

cm two-dimensional detector. Data on SANS intensity were col-

lected with a Q-range from 0.001 A21 to 0.4 A21 at 258C. Q is the

scattering vector and is related to the wavelength k and the scatter-

ing angle 2h by

Q ¼ 4pksinðhÞ ð3Þ

The instrument has pinhole geometry. Scattering intensities were

normalized using direct beam transmission measurements and were

reduced according to published protocols.24,25 Processing of the

data taken at different scattering lengths was performed using the

IGOR 6.2/IRENA software26 to obtain structural characteristics at

the level of fiber building and packing. To estimate the mesh size of

the cross-linking networks in the samples, the Debye-Bueche

model27 was used in the following form

IðQÞ / l3c

ð1þ Q2l2c Þ2ð4Þ

where lc is the correlation length. The correlation length of a net-

work is a measure of the spatial extent of the cross-linking regions

and reflects the average mesh size. A larger correlation length value

correlates with a larger average mesh size.28

Mass fractal dimensions were found using the fractal model (Dr.

A. Allen, NIST) implemented in IRENA and described in detail

within the program. Fractal analysis is often used to analyze materi-

als that have a repetitive unit which is appropriate for our polysac-

charide-based systems (see Supporting Information for the struc-

tures of chitosan, alginate and chondroitin). Fractal analysis is done

in the Porod (or high-Q) region of the I(Q) vs. Q plot. This region

corresponds to a range of distances smaller than the size of the scat-

tering objects so that the scattered neutrons are probing the local

structure of the repetitive unit. The fractal dimension (d) in mass-

fractal analysis is a number ranging from 1 to 3 which characterizes

the structure of the repetitive unit. For instance, a mass-fractal value

of � 1.7 corresponds to a polymer in good solvent whereas a value

of 2 or greater corresponds to a degree of branching.33 Scattering

from a mass-fractal is given as

IðQÞ / BQ�d ð5Þ

where d is the fractal dimension (obtained from the slope of the

LogI(Q) vs. LogQ plot, see Supporting Information) and B is the

prefactor in the power law (5) is indicative of the dimensional char-

acteristics of the mass fractal and/or its degree of swollenness.

Characteristics of individual fibers were acquired with the

ATSAS software.29 The radius of gyration of the cross-section (Rc)

was determined by calculating the pair distance distribution func-

tion of the fiber cross-section (Pc(r)) using indirect Fourier trans-

form methods in GNOM. The radius of gyration of the cross-sec-

tion describes the average distance of all area elements of the cross-

section from the center of scattering density. The r value at Pc(r) 50 gives the maximum linear dimension for the cross-section of the

scattering particle, dmax. The radius of gyration of the cross-section

of the scattering particle, Rc, is derived from the second moment of

Pc(r).

PcðrÞ ¼ 1

2p2

ZQIðQÞ � r sinðQ � rÞdQ ð6Þ

R2c ¼

R dmax

0PcðrÞr2dr

2R dmax

0PcðrÞdr

ð7Þ

Since the scattering intensity is directly proportional to the con-

centration (in mg/mL) and the molecular weight (in Da) of the con-

stitutive molecules, to normalize pair-wise distribution functions of

the cross-section, Pc(r), data for each polymer sample were divided

by the sum:

Xi

Ci�Mi ð8Þ

where i is the number of polysaccharide components Ci is the con-

centration of corresponding component (in mg/mL) and Mi is the

average molecular weight of the i-th polysaccharide (in Da).

Statistical AnalysisFive experiments were performed per sample for each mechanical,

porosity and density test. Six SEM images from each sample were

taken for analysis of the average pore size over the entire sample sur-

face. The Tukey-Kramer method was used to determine significant

differences between the average pore sizes of different sample sets.

One set of SANS data was obtained for structural analysis. SANS

analysis was performed on single samples. A Student’s unpaired t

844 Hyland et al.

Biopolymers

test or analysis of variance (ANOVA) was carried out to determine

the statistical significance (P\ 0.05) of differences in material me-

chanical properties, porosity and density.

RESULTS AND DISCUSSION

Mechanical TestingCompressive Testing. Shown in Figure 3, the elastic moduli

for hydrated samples increased with increasing polysaccha-

ride concentration. There was a statistically significant (P\0.01) difference between type C elastic moduli at 5% concen-

trations compared with type C at 2 and 3% concentrations,

demonstrating network stiffness is affected by polysaccharide

concentration for type C networks. Type A and B networks

also demonstrate a trend of elastic modulus increase with

polysaccharide concentration increase. All type A moduli

had statistically lower values than comparative type B and C

samples (P\ 0.01). The presence of chondroitin seemed to

improve material stiffness since the type A samples did not

contain chondroitin. Additionally, samples 3C and 5C had

statistically higher elastic moduli (P\ 0.01) compared with

samples 3B and 5B. It appears the order of component addi-

tion only significantly affects the elastic moduli for the two

higher concentration samples. This result may be due to

incomplete penetration of chondroitin into the polymer net-

work at higher polysaccharide concentrations if chondroitin

is added as the 2nd modifier. The inability of chondroitin to

diffuse freely may limit electrostatically driven chondroitin-

chitosan interactions, which could affect mechanical strength

of the networks. Unlike chondroitin, Ca21 may be able to

overcome steric hindrance because of its much smaller size.

Furthermore, it was reported that Ca21 diffusion in higher

concentrations of alginate likely increases the number of

cross-linking events which improved mechanical strength of

alginate hydrogels.30 For comparison with reported values,

dry 5C samples were compressed as well, with an elastic

modulus of 4.4 6 0.52 MPa (Supporting Information, Table

I), giving a significantly larger modulus than the largest pre-

viously obtained result (2.56 6 0.41 MPa)11.

Tensile Testing. As polysaccharide concentration of hydrated

samples increased, network tensile strength increased in gen-

eral (Figure 4). Type C samples had the greatest ultimate ten-

sile strengths, ranging from 1.8 kPa to 3.2 kPa while type A

and B samples were significantly less (P\ 0.01). Type B sam-

ples were either statistically similar or slightly stronger than

type A samples in terms of tensile strength. Therefore, the

addition order of the 2 modifiers is just as important for ten-

sile strength as it is for the compressive strength of the net-

works. The ultimate tensile strength for dry 5C samples was

71.2 6 4.6 kPa (Supporting Information, Table I) which is

about 22 times larger than the hydrated tensile strength for

5C.

In summary, mechanical testing shows that, as a modifier,

chondoritin can indeed significantly strengthen chitosan/al-

ginate networks, provided chondroitin is added before Ca21,

the other modifier.

FIGURE 3 Elastic modulus of each hydrated sample type. As total

polysaccharide concentration increased, elastic modulus also

increased. Samples are identified by mixing order (A, B or C) and

by initial mass fractions of chitosan and alginate used (2%, 3% or

5%). Mechanical testing results are presented as the average of five

sample tests with the standard deviation reported as the error. The

error bars correspond to one standard deviation. Such applies to

Figures 3, 6 and 7 as well.

FIGURE 4 Ultimate tensile strength of each hydrated sample

type. Type C samples had statistically larger tensile strength values

than type B samples possibly due to lack of chondroitin diffusion in

type B samples.


Biopolymers

Material Pore Size, Porosity, and Density

Highly porous and interconnected pore structures are needed

to ensure an environment conducive to cell proliferation and

attachment in addition to allowing the free flow of nutrients.

SEM images (Figure 5) suggest material pore sizes are gener-

ally very heterogeneous. In Figure 6, the histograms also

show pore size heterogeneity. To determine whether the aver-

age pore sizes for each sample type (A, B, C, 2, 3, and 5)

were statistically significant from each other, the Tukey-

Kramer method was used. At 95% simultaneous confidence

levels, average pore size for sample type B was greater than A

and C, and sample types A and C were statistically equiva-

lent. Lack of chondroitin penetration may have induced the

fusion of pores during the second freezing event, creating

slightly larger pores in type B samples. At 95% simultaneous

confidence levels, average pore size of sample type 3 was

greater than 2 and 5, and sample types 2 and 5 were statisti-

cally equivalent. Larger pores in type 3 samples may have

been due to the diffusion of polysaccharides prior to the first

freezing. Type 5 samples contained more total polysaccharide

content and diffusion may have been slow, resulting in

slightly smaller pores. Conversely, type 2 sample polysaccha-

rides could interact freely and form more complex networks

consisting of slightly smaller pores. Chung et al. observed a

FIGURE 5 SEM images of the nine sample types. Accelerating potential 1.0 kV, 30.7 mm 3 30

mm. Sample images depict the heterogeneous nature of the pores. Images are identified by mixing

order (A, B or C) and by initial mass fractions of chitosan and alginate used (2%, 3% or 5%) for

the purpose of this paper.

FIGURE 6 Pore size distributions for each sample. Average pore

diameter\D[ is reported for each distribution.

846 Hyland et al.

Biopolymers

similar heterogeneous pore population for chitosan-alginate

networks frozen at 2208C.12 Regardless of the mechanism,

the data show that there is an optimal polysaccharide con-

centration in terms of pore size.

In general, increased material porosity (Figure 7) corre-

lates with smaller starting concentrations of polysaccharide.

Samples that were made with 2% polysaccharide have an av-

erage porosity about 15% higher than samples made with

5% polysaccharide. As for adding chondroitin as an addi-

tional modifier, the general trend is that it leads to a decrease

in porosity as type B and type C samples are slightly less po-

rous than type A samples. Type C samples were also slightly

less porous than type B samples. More complete chondroitin

incorporation in type B samples may be the reason for this

effect. Porosities and pore sizes of these networks are similar

to other chitosan-alginate networks.11,12,20,31

As expected, density shows the opposite trend of porosity

as higher density correlates with larger starting concentra-

tions of polysaccharide (Figure 8); samples that were made

with 5% polysaccharide have an average density over two

times larger than samples made with 2% polysaccharide. As

for adding chondroitin as an additional modifier, the general

trend is that it leads to an increase in density as type B and

type C samples have higher densities than type A samples.

However, the order of chondroitin addition does not seem to

affect material density as type B and type C samples have

statistically equivalent densities.

To recapitulate, it seems that when chondroitin is added

before Ca21, it increases the mechanical strength and reduces

the average pore size and porosity, in comparison to when

Ca21 is added before chondroitin. However, the addition

order has no statistically significant effect on the density of

the material.

SANS Structural Analysis

Analysis of the SANS data was performed in an effort to

understand how the structural differences between polysac-

charide networks in solution at the nanoscale level and at the

level of individual fibers translate into the bulk material

properties after freeze-drying. Different SANS parameters

characterize different individual properties of the fibers or

the networks, however, taken together they might form a

consistent picture of structure-property relationships.

As a rule, the scattering intensity profile I(Q) vs. Q char-

acterizes the mass and/or volume of the scattering particle.

The larger the mass and/or volume, the greater the intensity

I(Q). In general, one might expect that networks comprised

of higher scattering particles, upon freeze-drying, will pro-

duce mechanically stronger materials. Another dimensional

parameter that describes the fiber cross-section is the radius

of gyration of the cross-section Rc, which is obtained from

pair-wise distance distribution function of the fiber cross-

section Pc(r). A larger Rc characterizes a greater cross-section

of the polysaccharide fiber, and thicker fibers are capable of

forming stronger materials when freeze-dried.

One of the important characteristics of the individual

fiber is the mass-fractal dimension d which defines the struc-

ture of the repetitive unit (building ‘‘brick’’) of the fiber. The

packing and compactness of this repetitive unit is character-

ized by mass-fractal prefactor B, which reflects the degree of

swollenness of the unit. Greater B values correspond to

greater swollenness of the polysaccharide fiber building unit,

and greater swollenness results in a weaker material after

freeze-drying. Correlation length or mesh size lc defines the

properties of the polysaccharide network, and smaller values

of lc are usually attributable to stronger networks. An illustra-

tive summary which compares the parameters examined in

FIGURE 8 Average density of each sample. Increased material

density correlates with larger starting concentrations of polysaccha-

ride.

FIGURE 7 Average porosity of each sample. Decreased material

porosity correlates with increased polysaccharide concentrations.


Biopolymers

these networks can be seen in Figure 9. Also, a pictorial ex-

planation of SANS parameters for the polysaccharide net-

works studied in this article is provided in the Supporting In-

formation, Figure S7.

The measured scattering intensity (Figure 10) for each of

the networks indicates greater scattering from mixed net-

works compared with pure alginate and pure chitosan sam-

ples. Increased scattering intensity describes the formation of

aggregates and is consistent with the development of fibril

networks, evidence that chitosan and alginate interactions

have occurred. Of the three mixtures, the scattering intensity

for the chitosan/alginate/Ca21 mixture was the smallest and

the chitosan/alginate/chondroitin mixture was the largest.

This is consistent with the formation of much larger, stronger

scattering assemblies in the presence of chondroitin which

are capable to reinforce the resulting material after freeze-

drying. Indeed, when freeze-dried, the scaffolds containing

chondroitin demonstrate the greatest mechanical strength

(Figures 3 and 4).

The correlation lengths (lc, Table I) for the five samples

showed that prior to mixing, the average mesh size for algi-

nate was much larger; its correlation length was larger than

the correlation lengths of chitosan and the three mixtures.

The correlation lengths for chitosan and the chitosan-algi-

nate mixture are quite similar. The average mesh size of algi-

nate thus decreases during the mixing process which suggests

the presence of chitosan-alginate interactions. Interestingly,

the correlation length of the Ca21-containing mixture was

smaller than the other two mixtures. These results may occur

due to alginate stiffening upon Ca21 addition, which was

known to shorten alginate chains. Stokke et al. also observed

a similar shortening evident from the relationship between

scattering intensity and Ca21 concentration in pure alginate

gels using small angle X-ray scattering (SAXS).32 Thus, due

to such contraction of the alginate polymer, the addition of

Ca21 created a more densely-packed system (smaller correla-

tion length). Smaller mesh-sized networks, in general, should

be expected to produce stronger bulk materials after freeze-

drying. Therefore, freeze-dried polysaccharide scaffolds

modified with Ca21 may demonstrate greater mechanical

strength. The correlation length or average mesh size for the

chondroitin-containing mixture was the greatest among the

three mixtures. A larger correlation length for the chondroi-

tin-containing mixture indicated that addition of chondroi-

FIGURE 9 A physical depiction illustrating the network parameters obtained from SANS

analysis.

848 Hyland et al.

Biopolymers

tin increased the average mesh size of the polysaccharide net-

work. Mesh size increase may be due to increased fiber thick-

ness upon chondroitin interaction with chitosan-alginate

fibers as well as electrostatic repulsion of the negatively

charged components. In a system where both these modifiers

are added, one might expect chondroitin to increase the fiber

thickness and Ca21 to condense and stiffen those fibers into

a stronger, compact system. When freeze-dried, this polysac-

charide scaffold with two modifiers shows the highest me-

chanical strength (Figures 3 and 4).

Mass-fractal dimensions, d, which define the structure of

the repetitive unit of the fiber for all five samples (Table I)

point to the formation of randomly-branched swollen poly-

mers (d values from 2.6 to 3.0).33 Additionally, the power-

law prefactor (B) from mass-fractal analysis reflects the

dimensions and/or the degree of swollenness of the repetitive

unit and is the smallest for the Ca21-containing samples (Ta-

ble I). Contraction of alginate upon Ca21 addition decreased

the swollenness (B) and after freeze-drying this sample may

demonstrate greater mechanical strength. The addition of

chondroitin also reduced the B value slightly compared with

the chitosan-alginate mixture. The decreased degree of swol-

lenness is also in agreement with the results showing chon-

droitin addition directly translates to the strengthening of

bulk mechanical properties on freeze-drying.

The radius of gyration of the cross-section (Rc), derived

from the analysis of pair-wise distance distribution function

of the cross-section (Pc(r)), can also be found in Table I.

Here, Rc is the contrast weighted average distance of all area

elements of the cross-section from the center of scattering

density and, in general, it characterizes maximum dimen-

sions of the fiber cross-section. The Rc value for the chitosan-

alginate network is larger than the Rc values of separate chi-

tosan and alginate fibers suggesting that upon mixing, chito-

san and alginate interact to form a thicker fiber, with a larger

cross-section than either chitosan or alginate alone. After

Ca21 addition, the Rc value of the chitosan-alginate network

becomes smaller, due to the contraction of alginate upon

interaction with Ca21. As mentioned above, such contraction

simultaneously leads to a more compact and more dense net-

work as evidenced by the decrease in mesh size lc and prefac-

tor B (Table I), thus suggesting a stronger bulk material upon

FIGURE 10 I(Q) vs. Q SANS profiles for multicomponent bio-

polymer networks: chitosan, cyan; alginate, black; chitosan 1 algi-

nate, orange; chitosan 1 alginate 1 CaCl2, violet; chitosan 1 algi-

nate 1 chondroitin, light green. Greater I(Q) values correspond to

larger scattering particles, e.g., the (chitosan1 alginate1 chondroi-

tin) mixture forms the biggest assemblies. Inset plot shows Guinier

plots for rod-like particles, lnQ*I(Q) vs. Q2, and the linearity in this

region confirms the formation of elongated fibers in all systems.

Color code on inset corresponds to main figure. Statistical error

bars correspond to one standard deviation and represent error in

the scattering intensity estimation. Error bars are large at the instru-

ment configuration overlap region but are smaller than the plotting

symbols at low Q.

FIGURE 11 Pair-wise distance distribution functions, Pc(r), for

the cross-section of the rod-like fibers of multicomponent networks:

chitosan, cyan; alginate, black; chitosan 1 alginate, orange; chitosan

1 alginate 1 CaCl2, violet; chitosan 1 alginate 1 chondroitin, light

green. Functions with two maxima are characteristic for the dumb-

bell shape of the cross-section. Value of r in A where Pc(r) goes to

zero defines the maximum dimension of the cross-section which for

all fibers is around 375 A.


Biopolymers

freeze-drying. In contrast, after chondroitin addition, the Rc

value for the chitosan-alginate network becomes larger due

to the incorporation of chondroitin into the chitosan-algi-

nate network and the thicker fiber that results. In Figure 11,

the pair-wise distance distribution functions of the cross-sec-

tion Pc(r) are plotted together. These functions reflect the

probabilities of finding different distances between two arbi-

trary points within the cross-section, and the area under the

curve characterizes the mass per unit length of the fiber. The

pattern of Pc(r) for all three mixtures corresponds to an

asymmetrical dumbbell shape of the cross-section, yet the

shape is more pronounced in the Ca21-containing mixture.

Fiber contraction upon Ca21 addition may be the reason for

this change in shape. Once again, we see that addition of

Ca21 causes fiber contraction while chondroitin addition

increases fiber thickness. Together, these modifiers can

increase the fiber density and therefore increase the network

bulk mechanical properties after freeze-drying.

To summarize, increased scattering intensity describes the

formation of aggregates and is consistent with the develop-

ment of fibril networks, evidence that chitosan and alginate

interactions have occurred. Additions of both modifiers indi-

vidually change the structure of chitosan-alginate networks

in different ways. Addition of Ca21 causes the contraction of

the network due to Ca21-alginate interactions. This contrac-

tion increased the stiffness of the fibers. Addition of chon-

droitin causes an increase in fiber thickness due to chondroi-

tin-chitosan-alginate interactions. Increased fiber thickness

results in greater material density which in turn may increase

material stiffness and strength.

CONCLUSIONSTissues such as cartilage, tendons, or ligaments exist in

mechanically demanding environments. To repair or replace

these materials, it is desirable to mimic their mechanical

strengths in engineered soft biomaterials. Creating the

strongest materials possible requires an understanding of

how individual network components and various conditions

affect bulk material properties. In the present work, we

examined how the addition of chondroitin affected the

properties of chitosan-alginate networks. Samples contain-

ing chondroitin were stiffer and had greater tensile strengths

than samples without chondroitin. However, the effective-

ness of chondroitin addition was dependent on the order in

which it was added. When added after the first lyophiliza-

tion (type B samples), chondroitin could not diffuse into

the networks. Therefore, type B samples were mechanically

weaker than samples where chondroitin was added prior to

the first lyophilization (type C samples). Effects of total

polysaccharide concentration were also studied. Higher con-

centrations were associated with greater mechanical

strengths. Porosity and density were notably concentration

dependent. Pore size was affected by both concentration and

order of chondroitin addition. Structural analysis of the net-

works complemented the findings in this paper. Correlation

length, dimensional characteristics of the repetitive unit,

and radius of gyration of the cross-section illustrated that

chondroitin addition increased fiber thickness while Ca21

addition caused fiber contraction thereby increasing fiber

stiffness. Together, the two modifiers improved network

density, resulting in greater stiffness and tensile strength.

This effort demonstrates the mechanical tunability and

enhancement of these materials for various tissue engineer-

ing applications.

All SEM images were collected with resources and assis-

tance from the Maryland Nanocenter. Thanks to Dr. W.

Chiou and Dr. L. Lai for SEM sample preparation and image

collection assistance; and Dr. W. Wang for helping with sta-

tistical analysis. The identification of commercial products

does not imply endorsement by the National Institute of

Standards and Technology nor does it imply that these are

the best for the purpose.

REFERENCES1. Baruch, L.; Machluf, M. Biopolymers 2006, 82, 570–579.

2. Suh, J. K.; Matthew, H. W. Biomaterials 2000, 21, 2589–2598.

3. Howling, G. I.; Dettmar, P. W.; Goddard, P. A.; Hampson, F. C.;

Dornish, M.; Wood, E. J. Biomaterials 2001, 22, 2959–2966.

4. Freier, T.; Koh, H. S.; Kazazian, K.; Shoichet, M. S. Biomaterials

2005, 26, 5872–5878.

5. Hu, Q.; Li, B.; Wang, M.; Shen, J. Biomaterials 2004, 25, 779–

785.

6. Shanmugasundaram, N.; Ravichandran, P.; Reddy, P. N.; Rama-

murty, N.; Pal, S.; Rao, K. P. Biomaterials 2001, 22, 1943–1951.

7. Park, D. J.; Choi, B. H.; Zhu, S. J.; Huh, J. Y.; Kim, B. Y.; Lee, S.

H. J Cranio-Maxfac Surg 2005, 33, 50–54.

8. Wang, L.; Shelton, R. M.; Cooper, P. R.; Lawson, M.; Triffitt, J.

T.; Barralet, J. E. Biomaterials 2003, 24, 3475–3481.

9. Vacanti, C. A.; Langer, R.; Schloo, B.; Vacanti, J. P. Plast

Reconstr Surg 1991, 88, 753–758.

10. Rees, D. A. Pure Appl. Chem. 1981, 53, 1–14.

11. Li, Z.; Ramay, H.; Hauch, K. D.; Xiao, D.; Zhang, M. Biomateri-

als 2005, 26, 3919–3928.

12. Chung, T. W.; Yang, J.; Akaike, T.; Cho, K. Y.; Nah, J. W.; Kim,

S. I.; Cho, C. S. Biomaterials 2002, 23, 2827–2834.

13. Lai, H. L.; Abu’Khalil, A.; Craig D. Q. M. Int J Pharm 2003, 251,

175–181.

14. Mikos, A. G.; Sarakinos, G.; Lyman, M. D.; Ingber, D. E.;

Vacanti, J. P.; Langer, R. Biotech Bioeng 1993, 42, 716–723.

15. Mooney, D. J.; Kaufmann, P. M.; Sano, K.; Mcnamara, K. M.;

Vacanti J. P.; Langer, R. Transplant Proc 1994, 26, 3425–3426.

16. Mooney, D. J.; Park, S.; Kaufmann, P. M.; Sano, K.; Mcnamara,

850 Hyland et al.

Biopolymers

K.; Vacanti, J. P.; Langer, R. J Biomed Mater Res 1995, 29, 959–

966.

17. Sirko, S.; von Holst, A.; Wizenmann, A.; Gotz, M.; Fraissner, A.

Development 2007, 134, 2727–2738.

18. Pieper, J. S.; Oosterhof, A.; Dijkstra, P. J.; Veerkamp, J. H.; van

Kuppelvelt, T. H. Biomaterials 1999, 20, 847–858.

19. Park, Y. J.; Lee, Y. M.; Lee, J. Y.; Seol, Y. J.; Chung, C. P.; Lee, S. J.

J Contr Release 2000, 67, 385–394.

20. Zhang, Y.; Zhang, M. J Biomed Mater Res 2001, 55, 304–312.

21. Berger, J.; Reist, M.; Mayer, J.; Felt, O.; Peppas, N.; Gurny, R. J

Pharm Biopharm 2004, 57, 19–34.

22. Abramoff, M. D.; Magelhaes, P. J.; Ram, S. J. Biophotonics Int

2004, 11, 36–42.

23. Glinka, C. J.; Barker, J. G.; Hammouda, B.; Krueger, S.; Moyer, J.

J.; Orts, W. J. J Appl Crystallogr 1998, 31, 430–445.

24. Kline, S. R. J Appl Chrystallogr 2006, 39, 895–900.

25. Whitten, A.; Trewhella, J. Small-Angle Scattering and Neutron

Contrast Variation. In Micro and Nano Technologies in Bioa-

nalysis; Humana, 2009, 307–322.

26. Ilavsky, J.; Jemian, P. J Appl Crystallogr 2009, 42, 347–353.

27. Debye, P.; Bueche, A. M. J Appl Phys 1949, 20, 518–525.

28. Soni, V. K.; Stein, R. S. Macromolecules 1990, 23, 5257–5265.

29. Svergun, D. I. J Appl Cryst 1992, 25, 495–503.

30. Kuo, C. K.; Ma, P. X. J Biomed Mater Res A 2007, 84, 899–907.

31. Ho, M. H.; Kuo, P. Y.; Hsieh, H. J.; Hsien, T. Y.; Hou, L. T.; Lai,

J. Y.; Wang, D. M. Biomaterials 2004, 25, 129–138.

32. Stokke, B. T.; Draget, K. I.; Smidsrod, O.; Yuguchi, Y.; Urakawa,

H.; Kajiwara, K. Macromolecules 2000, 33, 1853–1863.

33. de Gennes, P. G. Scaling Concepts in Polymer Physics; Cornell

University Press: Ithaca & London, 1979.

Reviewing Editor: C. Allen Bush


Biopolymers

Mutually Reinforced Multicomponent Polysaccharide Networks...Mutually Reinforced Multicomponent Polysaccharide Networks Laura L. Hyland,1 Marc B. Taraban,1 Boualem Hammouda,2 Y. Bruce

Documents